Data-driven Groundwater Depth and Risk Forecasting in the Central Sands Region of WI for Sustainable Management

Home / Research / Data-driven Groundwater Depth and Risk Forecasting in the Central Sands Region of WI for Sustainable Management
Project Number:

WR21R003

Other Project Number:

22-WSP-01-HUA

Funding Year:

2021

Contract Period:

-

Funding Source:

UWS

Investigator(s):
PIs:
  • Jingyi Huang, University of Wisconsin-Madison, Department of Soil Science
  • Ankur Desai, University of Wisconsin-Madison, Department of Atmospheric, Oceanic, and Space Sciences
Abstract:

Successful management of groundwater (GW) quantity and quality requires an improved understanding of controls of natural and anthropogenic factors on GW dynamics and developing models and decision-support tools for GW depth forecasting and management. Modeling GW depth dynamics is difficult and computationally slow because the water cycle is a complex process. The overall objective of this project is to combine high-resolution remote sensing data (e.g. precipitation, temperature, evapotranspiration, soil moisture, land cover and crop types), GW withdrawal, irrigation records and GW depth measurements to hindcast and forecast GW depth and nitrate health risk dynamics in Wisconsin Central Sands (WCS) using a data-driven machine learning (ML) framework. The project will evaluate different ML algorithms in their ability to hindcast GW depth at 20 USGS wells in the WCS on a monthly basis from 1958 to 2020 and perform scenario analysis to forecast GW depth and nitrate health risk from 2020 to 2050. The team will investigate controls and sensitivity of climate variability, soil moisture, land cover and crop types, water withdrawal, irrigation management, topography, geology and soil hydraulic properties on GW depth fluctuations. The team will present modeled and forecasted GW depth from 1958 to 2020 and GW depth and nitrate risk from 2020 to 2050 at state and national meetings and conferences to stakeholders, government agencies, and general public. The team will work with other researchers (e.g. Wisconsin DNR) and stakeholders to further improve GW monitoring, modeling, and forecasting using downscaled soil moisture and ET maps developed by our team and other environmental datasets. The team will build a Google-map like web platform with clickable features to display maps of modeled GW depth and nitrate risk across the WCS. These outreach activities will provide guidance for the stakeholders in Wisconsin for decision-making on groundwater management under a changing climate.

Project Reports: